Generally described, monitoring systems are used to obtain information from a variety of elements. In a representative use, a security monitoring system may provide real-time data reflecting the current status of a monitored environment, such as a physical location within a given premises. For example, a security guard may use a computer terminal to obtain video data from a number of cameras in order to assess the status of a premises. Accordingly, many conventional security monitoring systems can assist users in evaluating whether a security breach, or other monitoring issue, has occurred (e.g., whether there is an unauthorized presence within the premises). Further, some security monitoring systems, upon the detection of a condition, are operable to initiate a number of reactive measures. For example, the monitoring system may alert an appropriate authority, or notify one or more authorized users.
Although the traditional security monitoring system obtains information regarding the status of various aspects of a monitored environment, such as the status of physical devices or the presence or location of individuals, the outputs from most traditional security monitoring network data are fundamentally reactive in nature. With reference to the previous example, if a security monitoring system obtains motion detection data from a monitored premises, the data output for the traditional security monitoring network is typically limited to a determination of whether motion occurred and whether the detected motion is authorized. Both of these outputs are reactive in nature. Similarly, if a security monitoring network obtains live video data, the data output for the traditional monitoring network will be a transmission of the incoming video data to a display terminal, or more reactive, the archival of the video data. Clearly, the traditional security monitoring network cannot predict when motion will be detected or what the contents of the video motion may be. Thus, most, if not all, monitoring networks, are designed for, and therefore limited to, reactive data processing.
Although it may not be possible to predict events, particularly those linked to human behavior, with total accuracy, there are a variety of situations in which one or more factors may be utilized to establish a likelihood of an event occurring. In some limited situations, a single inputted factor, or condition, may have a sufficiently strong association with a target event such that the presence of the factor will likely determine whether the target event will occur. More commonly, however, the presence of a number of inputted factors, which if considered in isolation would have a limited association with a target event, may cumulatively indicate the likelihood of the target event occurring.
As applied to security monitoring networks and security processing services, the processing of data for the purpose of to the identifying an individual, such as facial recognition, fingerprint, retinal scan, and the like, may be useful for assessing security threats when the data used to identify an individual is linked to data linking that individual to a potential threat, based upon past behavior or other known risk factors In many instances, the risk factors may be unrelated to a specific individual, such as a state of alert at a premises. There are a number of situations in which the processing of multiple data inputs to produce a predictive threat assessment, that is, to process multiple data inputs to assess the likelihood of a target event, is clearly beneficial.
With reference to a security monitoring system implementation, there is an undeniable benefit from generating a threat assessment based on processing a wide variety of factors. For example, considered in isolation, the purchasing of a one-way airline ticket may not pose a sufficient threat to require additional investigation on behalf of law enforcement authorities. However, if the one-way ticket purchase is considered in conjunction with information, such as the purchase of the ticket with cash or the absence of checked baggage, the cumulative information could generate a threat assessment requiring at least some additional follow up, such as an automated notification to a perform an search at a security checkpoint. With another potential embodiment, a storeowner may wish to generate a reward assessment based upon predicted consumer's actions. In these situations, and many others, the data inputs can be interpreted to facilitate future actions. In each of these examples, however, because the assessment is not reactive, conventional monitoring systems are not well suited to provide such services.